Behind the numbers

India's policymakers need more timely unemployment data

Last week, the National Statistical Office released the Annual Report of its Periodic Labour Force Survey (PLFS), covering the period between July 2019 and June 2020. This period included the second quarter of the calendar year 2020, when the Indian economy was reeling from a stringent lockdown meant to contain the spread of Covid-19. Yet the difficulties that the pandemic and consequent lockdown imposed upon the broader Indian workforce did not appear, at first glance, to be adequately reflected in the PLFS. The headline that was widely reported was that the 2019-20 PLFS overall saw a reduction, rather than an increase, in unemployment, in that it fell from 5.8 per cent in 2018-19 to 4.8 per cent in 2019-20 (both years defined as being from July to June). It is worth noting that in general unemployment rates in subsistence conditions, as in much of India, do not always make sense. At low levels of income, unemployment is a luxury — people who need a daily wage to survive cannot spend a long time out of, or looking for, work. Even so, however, the unemployment numbers deserve closer scrutiny in order to ensure that the PLFS itself retains usefulness in its picture of the economy.

When scrutinised more closely, a pattern emerges in the PLFS data. For one, the survey reports two different estimates of employment and unemployment, one based on a worker’s “usual status”, or her ability to find work over the past year, and one based on her “current weekly status”, based on whether adequate employment was available to her in the week prior to the question being asked. It is unemployment under the usual status paradigm that has declined; under the current weekly status paradigm, unemployment stayed at the relatively high level of 8.8 per cent in all the years between 2017-18 and 2019-20.

The dynamics of a near-subsistence workforce under the pressure created by bad times are also very similar to those that are predicted by economic theory. As argued above, if unemployment is a luxury, then broadly consistent headline numbers will conceal some churn as individuals who cannot afford to stay home without work take on less appealing or less remunerative work. There is some indication of this happening in the PLFS. For example, among rural female workers, the “helper in household enterprise” category, jumped in 2019-20 by over 5 percentage points to 42.3 per cent. In general, economists are happier looking therefore at wage rates in subsistence economies rather than the overall unemployment rate in order to develop a clearer picture of the state of labour distress. The PLFS data on average earnings per day does indeed seem to confirm the notion that the lockdown reduced incomes sharply, especially for casual labour — wage rates increased at a steady rate over the three previous quarters of the year 2019-2020 (measured from July to June) before falling sharply in the April-June 2020 quarter.

The basic problem, however, from a policy point of view is that while the PLFS does provide valuable information about the state of the workforce, it comes too late to make a tangible difference to policy. There is no clear reason why the Indian statistical establishment cannot focus at least on a narrow and revealing set of indicators — average gross earnings rates, for example — and provide higher-frequency data that can properly inform policy. The lack of such data was keenly felt during the lockdown last year. If available, it might have energised the government into taking more effective and targeted welfare measures.